So far in the first two primer posts I’ve described some over-arching concepts about the data that goes into IDMP, how it’s used, and where it comes from. Let’s start piecing things together and understand what an IDMP message would look like.
I promise not to use XML for now – the HL7 Structured Product Language is not especially helpful at letting you see how things are put together, but I’d like to show what connects to what, and, and more importantly what doesn’t connect.
First, it’s important to understand that there are layers of product. I’m going to compare a lot of this to how things are currently done with the US Food and Drug Administration, since things there are similar, and IDMP hasn’t actually been implemented yet.
I will be using examples found in DailyMed using brand names so you can explore them yourself. No endorsement of products is implied here, and I don’t think any of the examples used below are current ACUTA customers, just products that show some of the trickier corners of the IDMP model.
- Medicinal Product: This is the root of the IDMP model (and its name: Identification of Medicinal Products). It’s the product as authorized (for sale or use in clinical trials), which may include – in some regions – multiple package sizes. It gets an MPID: Medicinal Product ID, which contains the country, the manufacturer code, and the product code (Note that the US National Drug Code [NDC] includes a package code, but not a country code).
- Packaged Product: This is the product as sold: a packaged product for retail, wholesale, professional samples etc. Each package has a PCID – the package ID, which is just the MPID with a suffix code for the specific pack size (making it very similar to the US NDC). Note that internal packaging may have its own PCID. Packages are usually grouped into batches that have a batch ID (BAID_1 for packages as-sold, BAID_2 for primary packages in contact with the product, if different), but I’m not going to spend time on that here today, especially as Europe will not be requesting batch IDs for their initial implementation.
- Pharmaceutical Product: This is the product as administered. Most products are administered as they are sold: prefilled single-use syringes, tablets, etc., but some are transformed. Some Medicinal Products may have multiple Pharmaceutical Products – the examples below include a cream and suppository in one package, and a set of tablets and capsules to be administered together – but each is a separate entity. IDMP has a special identifier for the Pharmaceutical Product ID (PhPID) in that it is calculated rather than issued by the agency. This lets products with the same active ingredients, dose form and strength share the same IDs, which will be very helpful in adverse event signal detection. IDMP requires that all the substances used in a Pharmaceutical Product be listed.
- Manufactured Items and Devices: When you open all the packages, this is what you see. Sometimes it’s the same ‘thing’ as the Pharmaceutical Product (e.g. tablets, pre-filled syringes), but on those cases where a product is combined or transformed, these are the sub-parts. When a device is used to administer the product (a spoon) or is integrated with the product (a syringe or inhaler), those are also described (although not in Europe’s first implementation plans). There are no specific IDs for Manufactured Items, but Devices may carry their own IDs (which are being standardized using UDI… but that’s another show). Manufactured items will also include lists of ingredients, which in rare circumstances include yet another kind of device that is part of the administered product, such as a scaffolding in a cell culture.
One of the first things we realized is that there is a lot of duplication between these levels, and that the Medicinal, Pharmaceutical and Manufactured items are often the same thing (or Medicinal is Pharmaceutical, or Pharmaceutical is Manufactured). ACUTA RIM (ARIM) is being engineered to minimize duplicate effort in data entry, so that if an item has to be described at one level, those details will apply to the other appropriate levels.
However, when you take a look at the IDMP data model, things aren’t as connected as you’d expect: Manufactured Items are related to the packages they’re in, not the Pharmaceutical Product. Because of that, although Manufactured Items’ ingredients should ‘add up’ to the Pharmaceutical Product, the data model doesn’t provide any way to look down the tree and sum them up automatically. This is a pretty rare case, though, where you would have two transformed pharmaceutical products in one kit – I haven’t found such an example.
Let’s start with something simple: Costco’s house brand, “Kirkland Signature Ibuprofen – ibuprofen tablet, film coated.” It has an NDC of 63981-604-90. In IDMP terms, it would have an MPID of something like US-63981-604, and a PCID of US-63981-604-90. It’s sold as a plastic bottle (no carton) of 500 tablets.
Except for the IDs, the diagram here focuses on the “what it is made up of” and leaves out the “what it is authorized for” part of the IDMP data. There’s a lot more information than this – manufacturers, storage conditions, etc., but today’s goal is just to understand the basic product concepts. In this case, it’s all pretty straightforward. The Medicinal Product (yellow) has an MPID. There’s a package (blue) with a manufactured item (tablets) in it made up of ingredients (orange); the pharmaceutical product (pink) has the same things. Note: The Pharmaceutical Product ID is simulated here with an abbreviated dose form code, the active ingredient, and the dose strength. In reality, it would be a secure hash of the codes for each of those items.
Let’s look at a more complicated product: Takeda’s PREVPAC, US NDC 64764-702-01, a combination of two antibiotics and a proton pump inhibitor for treatment of ulcer. There are two capsules of one type, four of another, and two tablets of the third in each daily dose blister, 14 blisters in a box. We have three pharmaceutical products, but still very simple manufactured items matching the pharmaceutical products.
Here we have three manufactured products, each of which corresponds to a pharmaceutical product. There are two layers of package, each of which has its own PCID. Note that the overall dose form is “Kit” – there’s no one thing that’s combined from the components.
For a last example, we have Sandoz’ Omnitrope, US NDC 0781-4004-36, for treatment of growth hormone deficiency. This is delivered as a vial of powder and a vial of diluent, eight each in a carton. It’s also sold as a cartridge that goes into a separately-sold pen, but we’ll leave that complication for another day.
Here, two manufactured items (vials of powder and diluent) are found each in a vial, eight of each in the carton, which has the only PCID. (the image appears to have clipped the quantity on the vials of diluent). This makes up a single pharmaceutical product. Note that there are four different dose forms present here:
- Powder – Manufactured Item
- Diluent – Manufactured Item
- Solution for Injection – Pharmaceutical Product
- Power and Diluent for Solution for Injection – Medicinal Product
Why are the Pharmaceutical and Medicinal Products listed differently? Because the Medicinal dose form is as sold, and the Pharmaceutical dose form is as administered.
Collecting and entering this information will be a daunting task. EMA is planning on assisting by transferring information in the current XEVMPD database for Article 57 into IDMP format, but certain data, such as packaging, are currently only a single descriptive field, and won’t be easily broken into the various pack sizes and manufactured items. On the flip side, though, the plans for ARIM will be to minimize the duplication you see above with repeated ingredient lists, as well as other properties of the Medicinal and Pharmaceutical Products and the Manufactured Items. Please contact ACUTA for how we can help you prepare for the IDMP requirements in Europe and elsewhere..